TextZoom

[ECCV2020] A super-resolution dataset of paired LR-HR scene text images

Zolly

[ICCV2023 oral] Zolly: Zoom Focal Length Correctly for Perspective-Distorted Human Mesh Reconstruction

torch3d_render

Total-text-segmentation

Cropped semantic segmentation label of scene text recognition dataset generated from total-text. Could be used to do recognition and semantic segmentation.